Sequence Relationship Between Long and Short Repetitive DNA Of

Home , Repeated sequence (DNA)

CORE Metadata, citation and similar papers at core.ac.uk Provided by Caltech Authors Proc. NaMl. Acad. Sci. USA Vol. 74, No. 10, pp. 4382-4386 October 1977 Biochemistry

Sequence relationship between long and short repetitive DNA of the rat: A preliminary report (long repetitive DNA/sequence homology/rat genome) JUNG-RUNG WU, WILLIAM R. PEARSON, JAMES W. POSAKONY, AND JAMES BONNER* Division of Biology, California Institute of Technology, Pasadena, California 91125 Contributed by James Bonner, August 8, 1977

ABSTRACT Long and short repetitive sequences of rat DNA fragments. Long repeated DNA has been used to drive whole can be isolated and characterized. Long [>1.5 kilobases (kb)] DNA tracers of various lengths to determine the interspersion sequences can be separated from short (0.2-0.4 kb) sequences by exclusion chromatography after renaturation of 4-kb DNA period of these sequences in long whole DNA. Our data are fragments to a repetitive Cot and digestion with the single- consistent with the model that short repeated sequences are strand-specific SI nuclease. (Cot is the initial concentration of present in the long repeated sequence elements in the rat ge- DNA in mol of nucleotides/liter multiplied by time in sec.) Long nome. repetitive DNA can be driven by an excess of whole rat DNA to measure its repetitive frequency. Excess long repetitive DNA can also be used to drive tracer quantities of either long (self- MATERIALS AND METHODS renaturation) or short repetitive DNA. Both the extent and the Preparation of DNA. Unlabeled DNA was extracted from rate of the renaturations are found to be similar, suggesting that rat ascites cells and labeled DNA was extracted from Novikoff long and short DNA fragments share sequences. When long DNA 3-4 kb long were prepared repetitive DNA is used to drive whole DNA tracers of various hepatoma cells.t fragments lengths, a 3.2-kb interspersion period is found. These data are by shearing DNA for 45 min at 7500 rpm in a Virtis 60 ho- consistent with the concept that short repetitive sequences are mogenizer (7) in 30 mM NaOAc (pH 6.8). DNA was sheared present within long repetitive DNA sequences in the rat ge- to 0.35 kb at 50,000 rpm in the Virtis 60 and in 66% glycerol (7). nome. The DNA was then passed over Chelex 100 (Bio-Rad), filtered, and precipitated with EtOH. Recent studies on repetitive DNA sequence organization Sizing DNA Fragments. Single-stranded DNA fragment (1-4, t) have revealed that there are two size classes of repeated lengths were determined by sedimentation through alkaline DNA sequences. In many organisms, interspersed sequences sucrose gradients. Isokinetic sucrose gradients (8) were formed 0.2-0.4 kb long comprise 50-70% of the repeated DNA while in SW41 tubes in 0.1 M NaOH using a Vmix of 10.4 ml, Cflask the remainder of the repeated sequences are substantially = 16.0% (wt/vol), and Cr, = 43% (wt/vol). Gradients were longer, more than 1.5 kb (1 kb = 1000 base pairs or nucleotides) centrifuged from 16 to 24 hr at 40,000 rpm. All tubes contained in length. The existence of two size classes of repeated sequences two markers of known molecular weight and samples were run poses a number of interesting questions. First, are long and short at least two times. Molecular weights were calculated from sequences kinetically different? This question is answered by sedimentation rates by the Studier (9) equations. measuring the repetition frequency and complexity of purified Preparation of Long Repeated DNA Fragments. DNA long and short repeated sequences. A second question is perhaps sheared to an average length of 4 kb was denatured and incu- more interesting: Do sequences that appear in short DNA se- bated at 65° in 0.3 M NaCI/10 mM piperazine-N,N'-bis(2- quences also appear in long sequences? This question is relevant ethanesulfonic acid) (Pipes) at pH 6.8 to an equivalent Cot of to the "integrator gene" hypothesis (5). Long "integrator" se- 5 using a factor of 2.31 for the reaction rate increase due to the quences also present throughout the genome as short inter- Na+ concentration. (Cot is the initial concentration of DNA in spersed sequences suggest batteries of control elements as mol of nucleotide/liter multiplied by time in sec.) After incu- proposed by Britten and Davidson (5, 6). bation, samples were diluted with an equal volume of 50 mM In this paper we present data concerning the kinetic pa- NaOAc/0.2 mM ZnSO4 at pH 4.2, and dithiothreitol was added rameters of long and short repeated DNA sequences and we to a final concentration of 5 mM. The final reaction mixture was examine the sequence relationships between the two classes of 0.15 M NaCl/5 mM Pipes/25 mM NaOAc/0.1 mM ZnSO4/5 sequences. Long and short repeated DNAs have been isolated mM dithiothreitol at pH 4.4. by nuclease digestion and agarose A-50 fractionation. Long DNA samples were incubated with S1 nuclease at 370 for 45 repeated DNA has been driven by whole DNA to determine min and the reaction mixture was chilled on ice and made repetition frequency, self-renatured to determine its com- 0.12 M in phosphate buffer. Duplex DNA strands were sepa- plexity, and used to drive short repeated DNA to examine rated by hydroxyapatite chromatography, eluted with 0.5 M cross-renaturation. We do not find any significant kinetic dif- phosphate buffer, and chromatographed on Bio-Gel agarose ference between the long repeated DNA elements and all re- A-50 (Bio-Rad) as described by Britten et al. (7). peated rat DNA sequences by these criteria. In addition, there The long unlabeled DNA used in the self-reaction and is evidence for some sequence homology between long and short long/short cross-renaturation and interspersion experiments repeated DNA sequences. A more sensitive experiment has been used to look for short sequences internal to long repeated DNA Abbreviations: kb, kilobase (1000 base pairs); Cot, initial concentration of DNA in mol of nucleotide/liter multiplied by time in sec; Pipes, The costs of publication of this article were defrayed in part by the piperazine-N,N'-bis(2-ethanesulfonic acid). payment of page charges. This article must therefore be hereby marked * To whom reprint requests should be addressed. "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate t Pearson, W. R., Wu, J. R., and Bonner, J. (1977) Biochemistry, in this fact. press. 4382 Biochemistry: Wu et al. Proc. Natl. Acad. Sci. USA 74 (1977) 4383

~=0.9 00

0 r0 250 50P (0

< ~~26 39 44

40 50 60 70 - 80 90 100 TEMPERATURE, 0C FIG. 2. Melting experiment of repetitive DNA duplexes of different fragment lengths. Repetitive DNA duplexes were isolated and fractionated as described in the legend of Fig. 1. Fractions 26, 39, and 44 were melted in a spectrophotometer equipped with a thermal cu- 10 20 30 40 50 60 70 FRACTION NUMBER vette. The temperature was raised at a rate of 0.50/min to 980. +, Fraction 44; 0, fraction 39; X, fraction 26; 0, native DNA refer- FIG. 1. Profile of rat repetitive DNA duplexes on agarose A-50. ence. DNA sheared to 4 kb was denatured and renatured to Cot of 5, digested with S1 nuclease, and bound to hydroxyapatite. The double- strand fraction (15%) was eluted with 0.5 M phosphate buffer and RESULTS chromatographed on Bio-Gel agarose A-50. The size of the fraction indicated was determined by alkaline sucrose sedimentation. The When 4-kb DNA is renatured to Cot of 5 and digested with SI fractions marked were used for the melting experiment in Fig. 2. nuclease and the duplexes are sized on agarose A-50, repetitive sequences are fractionated into two classes.t A typical column was isolated from 10 mg of 4-kb DNA. The DNA was denatured profile is shown in Fig. 1. Long (4 kb) fragments were used to for 5 min at 100°, incubated to Cot of 5 (13 min), and digested minimize creation of short fragments by random shear and with 250,ul of an SI nuclease preparation (the gift of Francine overlap of long repeated sequences. The incubation was carried Eden). This nuclease preparation has been extensively char- out to Cot of 5 to prevent renaturation of single copy sequences. acterized (10). The concentration used (25,ul/mg) corresponds (Long fragments have a higher effective Cot because of the to a digestion estimate of 0.85 which is 1.7 times the standard fragment length.) The long repetitive DNA excluded from incubation. After hydroxyapatite chromatography, 17% of the agarose A-50 is an average of 1.5-2.0 kb long. The short in- DNA was found duplexed (30 A20 units), and this duplex DNA cluded repetitive DNA is 0.2-0.5 kb long. was passed over agarose A-50. Fifty-five percent of the DNA To make certain that long duplexes did not contain single- was excluded. strand tails, we melted duplexes fractionated on A-50. A sample The 3H-labeled DNA used as tracer in the repetition fre- melting experiment of three fractions indicated in Fig. 1 is quency and cross-renaturation experiments was prepared as shown in Fig. 2. All fractions show more than 90% of native above. In this preparation, the enzyme-to-DNA ratio was 25 hyperchromicity. The melting temperatures range from 2.50 native for the material to 150 below Ail/mg; 19% of the DNA was bound to hydroxyapatite, and 47% below long (excluded) of the repetitive duplexes were excluded from agarose. native for the short (0.25 kb) fragments. Melting. DNA samples were melted in 0.12 M phosphate Renaturation of Selected Repetitive Sequences with Ex- buffer in a Gilford model 2400 spectrophotometer equipped cess of Whole DNA. 3H-Labeled 0.35-kb fragments were re- with a model 2527 thermal cuvette. Samples were melted at a natured to Cot of 100 and the duplexed repetitive sequences rate of 0.5°/min and the A260 was automatically sampled at were separated on hydroxyapatite. This total repetitive fraction 0.50 intervals. Hyperchromicity was calculated from the for- was then renatured with a 100- to 1000-fold excess of 0.35 kb mula of whole genomal DNA (repetitive plus single copy). The data are shown in Fig. 3A. The line drawn through the data displays - A26o(60') the best least-squares fit. The dashed line above the data is the H = A26o(980) fit for whole genomal rat DNA.t Table 1 (part A) shows the A26o(980) results of a least-squares fit to the data. About half of the re- after subtraction of the buffer absorbance at each tempera- petitive reaction takes place with a C0t1/2 of 0.337, corre- ture. sponding to a repetition frequency of 7400. The remainder DNA-DNA Renaturation. Samples that were not to be di- renatures with a C0t1/2 of 242, corresponding to a repetition gested by S1 nuclease were incubated in 0.12 M phosphate frequency of 10. buffer at 600 or in 0.48 M phosphate buffer at 700. After in- Labeled long repetitive DNA elements (those excluded from cubation, samples were frozen in dry ice/ethanol. Samples were A-50 agarose) were sheared to 0.35 kb and driven by 0.35-kb thawed and diluted to 0.12 or 0.14 M phosphate buffer and fragments of whole DNA to determine the repetition frequency passed over hydroxyapatite at 60°. The fraction bound was of the long repetitive sequences. Fig. 3B shows the renaturation eluted after thermal denaturation at 1000. The fraction and rate curve for the long repetitive sequences. Table 1 (part B) shows parameters for the renaturation curves were calculated from the results of a least-squares fit to the data. A large fraction of a nonlinear least-squares program (11). the DNA (60%) renatures by Cot of 10 with a Coti/2 of 0.153. 4384 Biochemistry: Wu et al. Proc. Nati. Acad. Sci. USA 74 (1977)

Table 1. Renaturation of isolated repetitive sequences with a 0.0 large excess of whole DNA Approximate 0.2 0 . repetition Component Fraction Rate Cot1/2 frequency N 0.4 A. Total repetitive DNA 1 0.45 2.97 0.337 7,400 9 0.6 2 0.23 0.00413 242 10 X Final fraction unreacted: 0.12 0 cr Goodness of fit: 3.4% o 0.8 I B. Long repetitive DNA 1 0.62 6.52 0.153 16,000 0m z 2 0.21 0.0140 71.4 35 o 0.0 Final fration unreacted: 0.15 B Goodness of fit: 2.2% z 0 z 0.2 0 estimate for the true fraction that repetitive sequences consti- tute at Cot of 5 is 16%, so the long sequences (45%), whether a: 0.4 F found only with a class of elements, should represent a 1/(0.45 LL X 0.16) = 14-fold enrichment of sequences over whole DNA. However, the enrichment found [Table 2 (part A)] is only 8-fold 0.6 _ [51.4 compared to 6.52 (Table 1 part B)]. Thus, the complexity found in the long repetitive DNA elements is twice that ex- on the basis of its fraction of the genome. This suggests 0.8 _ pected

0.0 1.0 0.001 0.01 0.1 1.0 10 100 1000 10,000 100,000 EQUIVALENT Cot 0.2 FIG. 3. Renaturation of selected repetitive sequences. (A) 3H- Labeled rat DNA (350 nucleotides) was renatured to Cot of 100 and the double-strand-containing fraction (30%) was separated on hy- 0.4 _ droxyapatite. This fraction was driven by unfractionated 350-nucleotide unlabeled rat DNA. The mass ratio of driver to tracer DNA was 100 from Cot of 0.001 to Cot of 0.5 and 1000 from Cot of 1.0 to Cot 0.6 of 1000. The line drawn through the data displays the best least- squares fit. The two repetitive components are also shown. The dashed line above the data is the whole rat DNA fit. (B) 3H-Labeled I- 0.8 F long repeated DNA fragments were isolated after fractionation on a. Bio-Gel agarose A-50 and sheared to 0.35 kb. These fragments were 0 O 1.0 driven by a 100-fold excess from Cot of0.01 to Cot of 0.1, by a 200- to x 500-fold excess from Cot of0.2 to Cot of0.5, by a 1000-fold excess from 0 0.0 Cot of 1.0 to Cot of 1000, and a 10,000-fold excess of unlabeled 0.35 whole rat DNA at Cot greater than 1000 in 0.12 M phosphate buffer 00 B 600. 0 at z 0.2 k z The remaining 20% of the DNA, which renatures by Cot of 0 zr 0.4 1000, contains less frequently repeated sequences, thus a lesser 0- repeat class. This part of the reaction may also include some single-copy DNA sequences. The measurements we have made on the repetition fre- 0.6 _ quency and complexity of long repetitive DNA in the rat show little difference between the long repetitive sequences isolated 0.8 _ from agarose A-SO and total (long plus short) repetitive sequences in whole DNA. The repetition frequency of long sequences is quite close to the repetition frequency of total re- 1.0 petitive sequences in whole DNA and the complexity of the 0.0001 0.001 0.01 0.1 1.0 10 100 1000 10,000 long size class is also about the complexity expected for the EQUIVALENT Cot fraction of sequences purified from whole DNA. FIG. 4. Renaturation of long and short repetitive sequences with Renaturation of Selected Repetitive Sequences with Ex- sheared long repetitive DNA fragments. (A) 3H-Labeled and unla- cess of Long Repetitive DNA. The complexity of the long beled long repetitive DNA was isolated as described in the legend of Fig. 3. The two DNA fractions were mixed and incubated to the Cot repetitive DNA elements was determined by their self-rena- values shown. (B) The unlabeled long repetitive fraction used in A turation kinetics, and the Cot curve is shown in Fig. 4A. Again was used to drive sheared short DNA fragments from the included the DNA preparation was sheared to 0.35 kb to exclude length peak in Fig. 1. The driver to tracer ratio was 10 for Cot of 0.002, 20 for effects. Table 2 (part A) presents the fit of the data. Our best Cot of 0.01, and 100 for Cot greater than 0.5. Biochemistry: Wu et al. Proc. Natl. Acad. Sci. USA 74 (1977) 4385 Table 2. Renaturation of long and short repetitive sequences by a large excess of long repetitive DNA Component Fraction Rate Cot1/2 A. Long repetitive sequences 1 0.64 51.4 0.019 2 0.13 0.161 6.2 Final fraction unreacted: 0.23 Goodness of fit: 1.0% R 4 0 / B. Short repetitive sequences 1 0.67 40.4 0.025 2 0.13 0.0768 13.0 0.2 0g Final fraction unreacted: 0.13 Goodness of fit: 0.5%

0 2000 4000 6000 8000 that some complexity of the long repeats may be shared with FRAGMENT LENGTH (nucleotides) that of the short repeats. fraction of rat DNA A more way to cross FIG. 5. The (R) containing long repetitive direct look for homology between long DNA elements as a function ofwhole DNA fragment length. Labeled and short repetitive sequences is to drive one preparation by whole rat DNA fragments of various sizes were prepared and driven the other. The results of an experiment in which sheared long to Cot of 0.5 (the lower Cot was used to adjust for the 10-fold lower repetitive DNA drove sheared short repetitive DNA are shown complexity ofthe repetitive driver) by the sheared long repeat DNA in Fig. 4B. The kinetics of the reaction are summarized in Table fraction used in Figs. 3 and 4. The solid line drawn to the data repre- 2 (part B). The rates of renaturation of long and of short re- sents the best least-squares fit to the data. The dashed lines are in- curves with the labeled tracers driven whole rat petitive sequences driven by a vast excess of repetitive terspersion by DNA long at Cot of 5 (middle curve) and at Cot of 50 (upper curve). The dotted DNA are the same within a factor of 2. Many sequences may line at bottom represents the background hydroxyapatite binding of be shared. long tracers in the absence of drivers. The fraction of the fragments Short sequences were not used to drive long sequences be- (F) containing duplexes was measured by binding to hydroxyapatite. cause of expected cross-homology. An unknown fraction of the The value plotted (R) is the value ofF corrected for the amount ofzero short DNA may be derived from the long elements by random time binding in the long tracers (Z). The values plotted are: R = (F mechanical shear. The kinetic complexity and crossreaction - Z)/(1.0 - Z). The values ofZ were calculated from a linear fit ofthe zero time binding data [Pearson, W. R., Wu, J. R., and Bonner, J. experiments are in fact sensitive to contamination of short re- (1977) Biochemistry, in press]. peated DNA with long repeated elements that have been sheared to short fragments. Such shear could happen in the initial preparation of the DNA or it could occur during the 30-40% of the short fragments would be driven by long se- hydroxyapatite separation of repeated from single-copy sequences. According to the results of Table 2 (part B), the sharing quences. Melting experiments (1, 12,t) suggest that some short of sequences between short and long is therefore more than repeated sequences are different from some sequences con- simple contamination. tained in long repeated elements. Short sequences exhibit more Interspersion of Long Repeated DNA Sequences. Long mismatch than do long sequences. But some short sequences repetitive sequences should be relatively free of short repetitive must be derived from long sequences, and the melt of the short sequence contamination. It is difficult to imagine a mechanical repetitive fragments shows a high precision component.t shear, renaturation, or nuclease artifact that would create long We cannot estimate how many of the short sequences are repeated sequences from short ones. We have used long re- derived from long sequences without knowing the in vivo peated DNA to drive various lengths of whole DNA tracers to length of the long repeated sequences in the genome. Some may determine whether the long repeated DNA is able to renature be longer than the 4-kb fragments used in our experiments. The the interspersed short repeats of whole DNA. Fig. 5 shows the length of the long duplexes excluded from agarose A-50 is about reaction of sheared long repeated DNA as driver at Cot of 0.5 1.5-2.0 kb. These may include some molecules that were me- with whole DNA tracers as a function of tracer length, corrected chanically sheared during hydroxyapatite fractionation. for zero time binding in the tracer. This experiment is similar It is possible to put an upper limit on the amount of con- to experiments done to determine the short repetitive sequence tamination by placing a lower limit on the number of short interspersion period in other organisms (2, 3, 13, 14,t). The sequences using the rat interspersed data.t At Cot of 5, 57% of dashed lines plotted in Fig. 5 are interspersion curves which can 2.5-kb DNA fragments contain a short repeated DNA sequence. be drawn through data from whole DNA tracers of different If the short repeated sequences are 0.25 kb long, 0.25/2.5 = 10% lengths driven by whole rat DNA at Cot of 5 and at Cot of 50,t of the bound DNA is repetitive, so 5.7% of the DNA contains but we have used a particular class of repeated sequences (Fig. short repetitive sequences renaturing by Cot of 5. At this Cot, 5, solid line) instead of whole DNA as driver. 0.16 of the genome is in true duplex, so 0.057/0.16 = 0.356 of The interspersion of long repeated elements is similar to that duplex sequences must be short. We find 50-60% of duplex of all repeated DNA sequences in the rat. The increase in hy- DNA included on agarose A-50, so 15-25% of that DNA or droxyapatite-bound DNA from 12% at zero length to 32% at 15%/50% = 30% to 25%/60% = 42% of the short duplexes may 1.8 kb cannot be due to a reaction between long repeated ele- be derived from long sequences. This is an upper limit. Many ments (sheared driver) and the complementary long segments short sequences are much longer than 0.25 kb and others may in whole DNA tracer. The difference in the single-stranded tail be interspersed at a longer period. If 33% of short sequences are length would only account for a 20% change (from 12.0 to derived from long sequences, short sequences would drive all 14.4%) if only long sequences (i.e., 1.5 kb) in a 1.8-kb whole long sequences in a cross-reaction experiment. Conversely, DNA tracers had renatured. Some of the sequences in the long 4386 Biochemistry: Wu et al. Proc. Natl. Acad. Sci. USA 74 (1977) repeated DNA preparation are apparently able to react with In sum, our evidence suggests that long and short repeated short sequences interspersed throughout the genome. DNA fragments share sequences in the rat. Possible contami- The 3.2-kb best-fit interspersion period is slightly longer than nant artifacts contribute to ambiguity as to the extent of this the 2.5-kb value found for whole short repeats in whole rat sharing, but we have not been able to provide any strong evi- DNA.t It suggests that some long repetitive DNA elements are dence that long and short repetitive sequence sets are different interspersed with a longer period than is found for the period from one another in the rat. of the short repeats. If long and short repetitive sequences are shared in the rat, a number of structural and organizational questions are raised. DISCUSSION Some short repeated sequences may be present as a single se- The renaturation experiments we have presented all suggest quence within a long repeated sequence or long repeated se- that there are no significant differences in complexity or rep- quences may be arranged as tandem arrays of short repeats. In etition frequency between most of the long and short repeated addition, some long repeated sequences may be made up of a DNA sequences in the rat. Eden et al. (16) have measured the number of different repeated sequences that are also found in repetition frequency, complexity, and sequence overlap be- the genome as short interspersed repeated sequences. Either tween long and short repeated sequences in the sea urchin. They structural model could provide the "integrator gene" function found that the repetition frequencies of the long and short re- suggested by Britten and Davidson (5). petitive sea urchin sequences in whole DNA are similar, with the long sequences repeated slightly less frequently. They also We thank Dr. Francine Eden and Ms. Denise Painchaud for the gift found the kinetic complexity of the short sequences to be about of the SI nuclease used in this study, Dr. Anthony Bakke for valuable 3 times that of the long DNA elements, reflecting the fraction discussion, and Mr. Bill Buchanan for his technical assistance. This of long sequences in the whole repetitive sequence population research was supported in part by U.S. Public Health Service Training of the sea urchin (about 30%). Grant GM00086 and in part by U.S. Public Health Service Research The cross-renaturation results-using long repetitive se- Grants GM13762 and GM20927. quences to drive short repetitive sequences-in the sea urchin are quite different from our results with rat DNA. In the sea 1. Davidson, E. H., Graham, D. E., Neufeld, B. R., Chamberlin, M. urchin, short repetitive tracer was driven at a rate 1/10 that of E., Amenson, C. S., Hough, B. R. & Britten, R. J. (1974) Cold the long repetitive driver. This result suggests that 10% or fewer Spring Harbor Symp. Quant. Biol. 38, 295-301. of the sequences in long repeated DNA elements can renature 2. Angerer, R. C., Davidson, E. H. & Britten, R. J. (1975) Cell 6, with short repeated DNA sequences. This result contrasts with 29-39. the results presented for the rat; we find no difference between 3. Efstratiadis, A., Crain, W. R., Britten, R. J., Davidson, E. H. & short repetitive tracer renaturation and the self-reaction of the Kafatos, F. (1976) Proc. Natl. Acad. Sci. USA 73,2289-2293. long repetitive driver. 4. Wu, J. R., Pearson, W. R., Wilkes, M. & Bonner, J. (1977) in The There are two possible explanations for these conflicting Molecular Biology of the Mammalian Genetic Apparatus, ed. in Ts'o, P.O.P. (Elsevier/North-Holland Biomedical Press, Am- results. First, the higher fraction of long repeated element sterdam, The Netherlands), Vol. 2, pp. 51-62. rat DNA may cause more cross-contamination problems. Sec- 5. Britten, R. J. & Davidson, E. H. (1969) Science 165,349-357. ond, the internal structure of the long repeated elements of the 6. Davidson, E. H. & Britten, R. J. (1973) Q. Rev. Biol. 48, 565- rat DNA may be different from those of sea urchin. 613. Under virtually identical digestion and fractionation con- 7. Britten, R. J., Graham, D. E. & Neufeld, B. R. (1974) in Methods ditions, we find that 40% to more than 50% of rat repeated DNA in Enzymology, eds. Grossman, L. & Moldave, K., (Academic sequences from 4-kb fragments are excluded on agarose A-50, Press, New York), Vol. 29, pp. 363-418. while Eden et al. (16) found with sea urchin that about 30% of 8. Noll, H. (1967) Nature 215,360-363. the sequences from 2-kb fragments were excluded. The cal- 9. Studier, F. W. (1965) J. Mol. Biol. 11, 373-390. we earlier that as much as 70% of 10. Britten, R. J., Graham, D. E., Eden, F. C., Painchaud, D. M. & culations presented suggest Davidson, E. H. (1976) J. Mol. Evol. 9, 1-23. rat repeated DNA may be longer than 1 kb. Britten et al. (10) 11. Pearson, W. R., Davidson, E. H. & Britten, R. J. (1977) Nucleic showed that almost 47% of isolated repeated sequence duplexes Acids Res. 4, 1727-1735. of sea urchin DNA are longer than 1 kb (mild S1 nuclease di- 12. Goldberg, R. B., Crain, W. R., Ruderman, J. V., More, G. P., gestion). The smaller the fraction of true short repeats in a ge- Barnett, J. R., Higgins, R. C., Gelfand, R. A., Galau, G. A., Britten, nome, the more difficult will be the isolation of the short repeat R. J. & Davidson, E. H. (1975) Chromosoma 51,225-251. fraction free of long repeat contamination. While it may be 13. Davidson, E. H., Hough, B. R., Amenson, C. S. & Britten, R. J. simple to show that two repetitive sequence populations are (1973) J. Mol. Biol. 77, 1-23. distinct, it is difficult to demonstrate unambiguously the degree 14. Graham, D. E., Neufeld, B. R., Davidson, E. H. & Britten, R. J. of true sequence overlap between them. (1974) Cell 1, 127-137. in or- 15. Schmid, C. W. & Deininger, P. L. (1975) Cell 6,345-358. Only a small fraction of the repeated DNA higher 16. Eden, F. E., Graham, D. E., Painchaud, D. M., Davidson, E. H. ganisms can be accounted for by sequences of known function. & Britten, R. J. (1977) Nucleic Acids Res. 4, 1553-1567. Some of the long repeated sequences must include the tran- 17. Brown, D. D. & Sugimoto, K. (1974) Cold Spring Harbor Symp. scribed multigene families, such as ribosomal and histone genes Quant. Biol. 38, 501-505. (17-19), and must be distinct from short repeated sequences. 18. Birnstiel, M., Telford, J., Weinberg, E. & Stafford, D. (1974) Proc. Long repeated DNA sequences also form higher precision Natl. Acad. Sci. USA 71, 2900-2904. duplexes than do most of the short repeated DNA (1, 10, 12, 19. Galau, G. A., Chamberlin, M. E., Hough, B. R., Britten, R. J. & Davidson, E. H. (1976) J. Mol. Evol. 200-224.